MI maps - muon2 scoring

[traininthedistance]

The change to protect the UCC pack in GR is expensive in erosity. Torie D had no macrocrop in  Kent and it had an erosity of 2 on the border between CD 2 and 3. The switch back to a macrochop ran the erosity on that border up to 14. Compare this chop in Kent to the one in Torie A which was also a macrochop, but was very rectangular and the border erosity was 9. This plan is more irregular than Torie A in Kent and the erosity is consistent with that.

This is a policy problem. By the way, I abandoned my thought that the cover rule only comes into play with whole county UCC severances. It won't work, as I will elucidate. The only way to protect UCC's adequately, is with a population based transference rule, or potentially a ban rule (maxi-pack rule is a requirement). Otherwise, UCC raids will tend to be rewarded, per the above, for example. When I have time (I don't know, buy hey I completed my assignments for the Columbia County DA, so I have my pro bono hours and then some, and can complete my application to become a NY lawyer), I will lay out the conundrums using Mike's "hideous" MI map, that chops the Detroit UCC to bits, while being rewarded for it.

Would you want this ban to include single-county UCCs?  Especially if you do, there will be cases where maxi-pack on all UCCs is physically impossible (such as, say, SEPA) and I presume the rules would then just only allow those maps with minimum UCC chops.  As mentioned in passing earlier, I especially anticipate trouble in Central Florida.

Also, as for rewarding for UCC chops, try comparing my most recent "serious" alternative (which is the same as Train B outstate, and completely respects UCC cover, though it still does not pack Grand Rapids), but adopts muon's 47% BVAP chop-minimzing Detroit).  That should ease your mind somewhat, once muon scores it?

As always, I tend to be more put off by cover violations than by pack violations, though the latter is non-ideal and I'm happy to have it dinged in the scoring.  I will once again point to western PA as a case where violating pack can lead to nicer lines without materially damaging the spirit of the UCCs.

[traininthedistance]

Alright, I've got an alternate chop-minimzer which doesn't have any traveling chops and, as a bonus, isn't so obviously ludicrous in the rural areas north of GR.  The downside is that inequality has to go up.

MI train 2015E






The Detroit area is *almost* the same; Sylvan Lake just SW of Pontiac was switched from 11 to 9.

The Jackson chop is, alas, a macrochop.  Tuscola and Washtenaw are not, though.  I have not rigorously tested if chopping Owosso instead would lower erosity.  Being non-macro chops, I imagine it oughtn't make much difference?

[traininthedistance]

train - I'll look at the score for your new offering when I get a chance. I'm curious if you can put up a better version that keeps both VRA CDs over 50% than my latest. With my three UCC penalty points in Detroit that should provide some room to work.

Yeah, you can shift things in Detroit to give the latest map two 50%+ districts, at the cost of one cut.

I also noticed that there was a stray precinct somewhere up north which would have pushed inequality to the absolute limit, so I took the opportunity to rejigger lines in districts 1-5.  I *think* this saves two erosity points anyway... so they can be easily grafted onto the other Detroit.

What this map sadly does not do is care much about inequality.  That number remains high and, with this arrangement, is staying high unless you add another chop, to give the northern five districts another 11K people or so.  So if inequality is a fully-equal leg on the Pareto stool, that would mean it's unlikely to knock many other plans out.  I've been thinking about trying to make an inequality-minimizing map (within reason), but would certainly go to a different template for that.  (Note that switching Sylvan Lake from 11 to 9 would help average inequality, but it hurts erosity and does nothing for the in extremis inequality measure currently in place.)

MI train 2015F

[traininthedistance]

The chop of Clinton is annoying since it incurs both the chop and UCC cover penalty. As I noted in my previous comments something has to chop into either Detroit or Lansing to get low inequality based on the SE MI regions. Seeing the effect of the chop of Clinton, I'm inclined to think that anything short of a macrochop of a UCC shouldn't get a cover penalty. This would only affect multi-county single-district UCCs, and put them on an even basis with single county UCCs.

I'm doubt I'd be on board with that change, as currently suggested.  I'd be more likely to either support adding cover penalties to single-county UCCs, or allowing non-macro chops in multi-county multi-district UCCs as well.  I don't think that we should be saying that inequality-reducing chops should have to go into Lansing rather than Detroit, but that's what your proposed rule seems to do. Or just live with the status quo– if that chop gives you enough of a inequality boost, it shouldn't be the end of the world that it counts twice.  

...

This is not likely to be a winning anything: it chops the Lansing UCC (something I had been avoiding for the longest time, and still would not particularly like to see rewarded... also not a huge fan of the Thumb split), for what's really only a marginal improvement in inequality (does not beat your latest map).  I'll continue to fiddle.

But, it continues to keep raw county cuts pretty minimized, so perhaps worth showing for that purpose. And Jackson is no longer a macrochop.

[traininthedistance]

This should be a (temporary) winner on CHOP + INEQUALITY.  It's not what I'd call an ideal map– severing Saginaw from Bay City/Midland still strikes me as far from ideal, and 3 is erose.  I'm sure this map will do nothing to challenge Torie's efforts to drive down the other pole.  But it gets things pretty low without incurring any UCC penalties (which should allay muon's concern wrt Clinton above), and with really not too many chops.  You'll note that one township in St. Joseph's County that was moved from 6 to 7... that can go right back to 6 and the map will still pass muster.  But making that microchop saves two inequality points, bringing average INEQUALITY down to 795, so it's worth it.

I suspect that more inequality-reducing chops will work on a one-to-one basis, so I'm not sure if you can push the frontier much further.  In particular, getting a UCC pack penalty by sending 10 into Lapeer rather than 11, and rearranging Detroit around that, gets us one point.  But I'll see if there's any other low-hanging fruit (like the 6 microchop) that I'm missing.









I will say that I'm not entirely sure that CHOP + INEQUALITY as a combo measure is my favorite way to get inequality into the scoring mix– though I'm also not sure what my preferred alternative would be, so happy to run with it as a hypothetical for now.

[traininthedistance]

I suspect that more inequality-reducing chops will work on a one-to-one basis, so I'm not sure if you can push the frontier much further.  In particular, getting a UCC pack penalty by sending 10 into Lapeer rather than 11, and rearranging Detroit around that, gets us one point.  But I'll see if there's any other low-hanging fruit (like the 6 microchop) that I'm missing.

I lied.  Microchopping into 4 (and chaining chops through 4->3->7->11) gets us another two-point boost on INEQUALITY. Average is down to 504.



(Detroit is the same except that 9 takes in Lake Angelus.)

We're really reaching the point of diminishing returns now.

I'm getting the impression that this is no place to discuss the partisan impact of these maps, no?

Muon's been scoring some of them on "skew" and "polarization", which he can explain better than I. While the whole point of this project is to try and create a set of rules that tries to take the partisan out of the redistricting, I for one think the "skew" measure is a very valuable and important check to make sure that these maps don't all turn out to be crypto-gerrymanders (which is certainly a potential danger with any "blind" rule-based system).  "Polarization" I'm less sanguine on, as it gives an incentive to create competitive districts and honestly some electorates are naturally polarized and I'm fine with a map which reflects that.

Obviously, with the current MI map being a Republican gerrymander, all of these proposals would be likely to help Dems to some extent.  But that's far from true for every state, and some of them don't help that much.